Ultrasound in Med. & Biol., Vol. 24, No. 5, pp.

739 749, 1998

Copyright 1998 World Federation for Ultrasound in Medicine & Biology
Printed in the USA. All rights reserved
0301-5629/98 $19.00 1 .00

PII S0301-5629(98)00034-9

Original Contribution
DISSOLUTION OF MULTICOMPONENT MICROBUBBLES IN THE
BLOODSTREAM: 1. THEORY
ALEXEY KABALNOV,* DAVID KLEIN, TIMOTHY PELURA, ERNEST SCHUTT and JEFFRY WEERS
Alliance Pharmaceutical Corporation, San Diego, CA
(Received 20 October 1997; in final form 2 March 1998)

AbstractThe problem of dissolution of a bubble in the bloodstream is examined. The bubble is assumed to be
filled with a mixture of a sparingly water-soluble gas (osmotic agent) and air. The dissolution of the bubble has
three definite stages. In Stage 1, the bubble quickly swells in air. The swelling ratio depends on the surface
tension, blood pressure, level of oxygen metabolism and initial mole fraction of osmotic agent in the bubble. In
Stage 2, the osmotic agent slowly diffuses out of the bubble. The squared radius decreases nearly linearly with
time, at a rate proportional to the Ostwald coefficient and diffusivity of the osmotic agent. In Stage 3, the partial
pressure of the osmotic agent becomes so high that it condenses into a liquid. In order to prolong the lifetime of
5-mm bubbles in the bloodstream from < 1 s (as found with pure air), the osmotic agent must have a low Ostwald
coefficient (< 1024) and a relatively high saturated vapor pressure at body temperature (> 0.3 atm 5 3 3 104
Pa). 1998 World Federation for Ultrasound in Medicine & Biology.
Key Words: Echocardiography, Ultrasound scattering, Ultrasound contrast media, Microbubbles, Microshells,
Microballoons, Ostwald coefficient, Dissolution kinetics, Molecular diffusion, Epstein-Plesset theory.

INTRODUCTION

where I is the scattered intensity, I 0 is the incident

intensity, n is the number density of scattering particles,
V is the scattering volume, k is the wave number, r is the
radius of the particle, g c is the compressibility term, g d
is the density term, u is the scattering angle equal to 180
for backscattering and d is the distance from scatterers.
The compressibility and density terms are determined by
the following equations:

There is a considerable interest in the development

of intravenously administered contrast media for ultrasound imaging. Since the pioneering study of Gramiak
and Shah (1968), who injected a shaken saline into aorta
and obtained a cloud of echoes due to the microbubbles
of air, many efforts has been made to create an ultrasound contrast medium based on microbubble dispersions (Balen et al. 1994; Goldberg et al. 1994). Currently
several companies are developing new microbubble media for the ultrasound contrast; for recent reviews, see
(Schlief 1996 and Fritzsch and Schlief 1996).
The main advantage of microbubbles, as compared
to other contrast media, is their high scattering intensity
per particle. The intensity of ultrasound scattered by
spherical particles that are much smaller than the wavelength of the ultrasound radiation is determined by the
formula (Morse and Ingard 1968):
I
1
k 4r 6~ g c 1 g d cos u ! 2
, nV
I0 9
d2

gc 5

ks 2 km
3rs 2 3rm
; gd 5
km
2rs 1 rm

(1a)

where k s and k m are the compressibilities of the scatterer

and the medium, and r s and r m are the densities of the
scatterer and the medium, respectively. The compressibility of gaseous particles is many orders of magnitude
higher than that of liquids and solids. Because of this, the
scattering efficiency from gaseous particles is many orders of magnitude higher than that from other objects.
For more complicated systems, such as air bubbles covered by viscoelastic shells, the scattering law becomes
more complicated and depends on the mechanical properties of the shell (de Jong 1993).
From eqn (1), it is evident that the particle size of
microbubbles for intravenous injection plays a signif-

(1)

Address correspondence to: Dr. Alexey Kabalnov, Alliance

Pharmaceutical Corporation, 3040 Science Park Road, San Diego, CA
92121, USA. e-mail: ekabalnova@aol.com
739

740

Ultrasound in Medicine and Biology

icant role and must be carefully selected. Indeed, large

bubbles are unable to pass through the capillaries,
whereas small bubbles are considerably less effective
scatterers of ultrasound due to the very steep I ; r6
dependence. It is now accepted that the size must be in
the range of 17 mm (Goldberg et al. 1994).
Early attempts to prepare microbubble dispersions
for ultrasound contrast were not successful because of
the very rapid clearance of microbubbles from the body
(Goldberg et al. 1994). The rapid clearance was explained by Meltzer et al. (1980), who suggested that air
microbubbles can dissolve in blood (Meltzer et al. 1980).
Indeed, the blood is saturated with air in the lungs under
atmospheric pressure, whereas the air pressure inside the
bubbles is larger than atmospheric due to Laplace pressure and blood pressure effects. An additional effect that
makes microbubble lifetime shorter is the oxygen metabolism by tissues (Yang 1971).
The kinetics of the dissolution process for spherical
bubbles in a liquid was discussed by Epstein and Plesset
(1950). Adapted for the dissolution of air bubbles in the
bloodstream, the theory predicts the following rate of
decrease of the particle radius with time:

dr
p# * 1 2 s /r 1
1
5 2DL
1
dt
p amt 1 4 s /3r r
p Dt

(2)

where D is the diffusivity of air in water, L is the

partition coefficient of air between water and gas phase,
(see below), p atm is the atmospheric pressure, p# * 5
4300 Pa is the excess pressure, which has a contribution
from both the systemic blood pressure and the oxygen
metabolism, s is the surface tension, r is the radius of the
bubble and t is the time. (The origin of the p# * and the
chosen numerical value is in detail discussed in Appendix.) The predicted lifetime for an air microbubble with
a radius of 2.5 mm is ; 0.17 s (Fig. 1).
The dimensionless ratio of the solubility of the gas
in the liquid to the gas density, which enters eqn (2), is
known in solution chemistry as the Ostwald coefficient,
viz:
L5

c water
.
C gas

Volume 24, Number 5, 1998

Fig. 1. Dissolution of air bubble in the bloodstream, as predicted by eqn (2). s 5 70 mN/m; p# * 5 4300 Pa, r 0 5 2.5
m m; L and D values are those of nitrogen (Table 1).

Conversely, gases of low polarity, such as aliphatic hydrocarbons and fluorocarbons, prefer to stay in the vapor
phase and their Ostwald coefficients are much less than
unity.
From eqn (2), it is clear that microbubbles filled
with gases having smaller values of L will live longer.
Currently, several groups are exploring the development of microbubble ultrasound contrast agents in
which the microbubbles are filled with low Ostwald
coefficient gases (Bernstein et al. 1997; Correas and
Quay 1996; Kabalnov et al. 1995; Lohrmann 1996;
Mattrey et al. 1994; Porter et al. 1996; Quay 1995;
Schneider et al. 1995; Schutt et al. 1997; Unger et al.
1994). The Ostwald coefficient is not the only parameter controlling the microbubble stability, however.
The other important factor is the saturated vapor pressure of the osmotic agent at body temperature. If the

(3)

The ratio represents the partition coefficient of a solute

between the gas phase and dilute aqueous solution (Fig.
2). The coefficient L is independent of concentration,
provided that the interactions between solute molecules
are weak (i.e., in the dilute region). Ostwald coefficients
are known for a wide variety of gases (Wilhelm et al.
1977). For polar gases, such as CO2 and NH3, the partition coefficient is strongly in favor of the water phase.

Fig. 2. Definition of Ostwald coefficient.

Microbubble ultrasound contrast A. KABALNOV et al.

vapor pressure is relatively low, the osmotic agent

may condense into liquid and the contrast will be
lost.
The application of the Epstein-Plesset formalism
to the gas exchange between the bubble and blood is
appropriate when the microbubble is filled with a
single gas. Thus, even in the case of air, the theory is
not strict due to the difference in Ostwald coefficients
and diffusivities of nitrogen and oxygen. This effect
becomes more significant when the filling gases differ
substantially in their Ostwald coefficients. Note that
even if the bubbles are filled with a single gas before
injection, they will swell in air and become multicomponent during the passage through the lungs postinjection.
Recently, Van Liew and Burkard (Burkard and
Van Liew 1994; Van Liew and Burkard 1995a,b)
examined the dissolution of bubbles of a sparingly
water-soluble gas (called below the osmotic agent) in
the bloodstream. The authors show that, after injection, the bubbles initially swell in air and then slowly
release the gas with a rate proportional to the Ostwald
coefficient and diffusivity of the osmotic agent. The
approach of the present article is similar to that of Van
Liew and Burkard. The main difference is that the
possibility that the osmotic agent may condense into a
liquid is taken into account in this report. This is an
important factor, since the low Ostwald coefficient for
a gas is typically achieved at the expense of a rather
low saturated vapor pressure.
The article has the following outline. First, the
problem of dissolution of two-component bubbles in
the bloodstream is examined. Air is modeled as a
pseudocomponent, and the consumption of oxygen by
tissues is completely neglected. The effects that are
accounted for are the Laplace pressure and systemic
blood pressure pblood. In the second part, an approximate analytic solution is provided for the case when
the Ostwald coefficient of the filling gas is much lower
than that of the air component (i.e., of both oxygen
and nitrogen). The formal treatment is similar to the
problem of Ostwald ripening in emulsions with a
two-component disperse phase (Kabalnov et al. 1987).
In the third part, air is no longer considered to be a
pseudocomponent, and the effect of the oxygen consumption/carbon dioxide production in tissues is taken
into account. Fortunately, the formalism of the previous section is also applicable in this case, with the
systemic blood pressure pblood being replaced by an
effective value p# *, which has contributions from
both the oxygen metabolism and systemic blood pressure.

741

STATEMENT OF PROBLEM AND

SIMULATIONS
Consider a microbubble of radius r, filled with two
gases, an osmotic agent and air. Air is modeled as a
pseudocomponent, and both air and the osmotic agent are
considered to be ideal gases. The osmotic agent and
air components are characterized by Ostwald coefficients L F ,, L A and water diffusivities D F and D A ,
respectively. Let C A (t) and C F (t) be the concentrations
of the gases in the bubble (in mol/m3). The condition of
the mechanical equilibrium of the bubble in the bloodstream states that:

~C F 1 C A! RT 5 p A 1 p F 5

2s
1 p blood 1 p atm
r

(4)

where p blood is the systemic blood pressure, p atm is the

atmospheric pressure, p A and p F are the partial pressures
of air and the osmotic agent in the bubble, respectively,
s is the surface tension, R is the gas constant and T is the
absolute temperature. It implies that the pressure of the
two gases in the bubble counterbalances the Laplace
pressure, blood pressure and atmospheric pressure. The
steady state1 diffusional flux of osmotic agent from the
bubble can be written as (Crank 1975):

JF 5 2

d 4p
~C Fr 3! 5 4 p rD F@c F~r! 2 c F~`!#
dt 3

(5)

where the lower case c F values are the concentrations in

water (in mol/m3). The concentration c F (r) is the subsurface concentration in water in direct contact with the
surface of the bubble. Under conditions of local equilibrium, c F (r) 5 L F C F . On the other hand, there is no
osmotic agent in the lungs, so, at infinity, c F (`) 5 0.
One concludes that:

d
~C r 3! 5 3rD FL FC F.
dt F

(6)

Similarly, for the diffusion of air:

d
p atm
~C Fr 3! 5 3rD AL A C A 2
.
dt
RT

(7)

1
In the analysis, the nonsteady-state effects are neglected. The
time of establishing the steady state for diffusion from r 5 2.5-mm
particles has an order of r 2 /D ; 6 ms and is an insignificant correction
for the systems of interest.

742

Ultrasound in Medicine and Biology

Volume 24, Number 5, 1998

Here, the condition that the blood is saturated with air at

infinity is imposed, so that c A (`) 5 L A p atm /RT. Introducing dimensionless parameters and variables:

m5

2s
p blood r
DA
;n5
; 5 r; d 5
;
p atmr 0
p atm r 0
DF

xA 5

C ART
C FRT
DF
; xF 5
; t 5 2 t;
p atm
p atm
r0

x Fr 3 5 F; and x Ar 3 5 A,

(8)

where r 0 is the initial particle radius r(t 5 0), one

obtains:
F 1 A 5 mr 2 1 ~1 1 n ! r 3

(9)

3L F
dF
52 2 F
dt
r

(10)

3dLA
dA
5 2 2 ~ A 2 r 3!.
dt
r

(11)

The initial conditions of the problem are:

r ~0! 5 1

(12)

A~0! 1 F~0! 5 m 1 n 1 1

(13)

F~0!
5 X F.
A~0! 1 F~0!

(14)

The first condition is obvious; the second states that, just

after injection of the bubbles into the bloodstream at t 5
0, the partial pressures of the gases counterbalance the
Laplace pressure, systemic blood pressure and atmospheric pressure. The third condition defines the mole
fraction of the osmotic agent in the gaseous mixture prior
to injection, X F . If the bubbles were initially filled with
osmotic agent only, X F 5 1; for bubbles filled with air
only, X F 5 0.
Figures 3A and 3B show the results of numerical
simulations of eqns (9) through (11) for an airn-perfluorobutane mixture with the following values of parameters and initial conditions: T 5 37C, n 5 0.043 (corresponding to p blood 5 4300 Pa), m 5 0.56 (corresponding
to s 5 70 mN/m), r 0 5 2.5 mm, at the initial mole
fractions X F 5 1 and 0.25. The Ostwald coefficient and
diffusivity of the air component are assumed to be those
of nitrogen; D F and L F of perfluorobutane are estimated
as described elsewhere (Kabalnov et al. 1990), see Table

Fig. 3. Dissolution of n-C4F10-filled bubble. s 5 70 mN/m;

p# * 5 4300 Pa; r 0 5 2.5 mm. (A) X F 5 1; (B) X F 5 0.25.

1 and Appendix. One may notice the following sequence

of events during the dissolution:
(1) At the first stage, the kinetics of dissolution are
characterized by a rapid swelling of the osmotic
agent bubbles in air (Fig. 3A), or conversely by a
rapid diffusion of air from the bubbles (Fig. 3B).
Whether the bubbles swell or shrink depends on the
initial partial pressure of air in the bubble: bubbles
with a partial pressure of air , 1 atm swell, while
those with a partial pressure . 1 atm, shrink.
(2) After this stage of rapid equilibration, the partial
pressure of air in the bubble establishes at 1 atm and
stays essentially constant during further dissolution.
The reason for this behavior is discussed in the

Microbubble ultrasound contrast A. KABALNOV et al.

743

Table 1. Parameters used in modeling.

Gas

Saturated
vapor
Boiling
Diffusion
Saturated vapor Saturated vapor concentration Water solubility
point
Molar volume
coefficient
pressure at 37C pressure at 25C (25C, mol/
of liquids
Ostwald coefficient
(C) (20C, cm3/mol) (D 3 10 9 , m2/s*)
(Pa)
(Pa)
m3)
(25C, mol/m3)
(L 3 10 6 )

N2
O2
C2F6
278.1
n-C4F10 22
n-C5F12
29
n-C6F14
59

105
153
177
201

19 (25C)
21 (25C)
8.6 (20C)
6.9 (20C)
6.3 (20C)
5.8 (20C)

3.5 3 106
3.8 3 105
1.3 3 105
4.8 3 104

2.8 3 106
2.6 3 105
8.5 3 104
2.9 3 104

1147
105.6
34.14
11.66

1.45
0.021
4 3 1023
2.7 3 1024

14480 (35C)
27730 (35C)
1272 (25C)
202 (25C)
117 (25C)
23 (25C)

* For oxygen and nitrogen, data from Lide (1991). For fluorocarbons, estimated from the molar volume with Hayduk-Laudie equation (Hayduk and
Laudie 1974). All the data refer to water as a solvent.

PCR Catalog (1992).

Lawson et al (1978); Kabalnov et al. (1990).

Reed (1964).

Kabalnov et al. (1990). For C2F6 and n-C4F10, approximation in the homologous series of linear perfluoroalkanes.

For oxygen and nitrogen, data from Wilhelm et al. (1977). For the other compounds, the ratio of the water and vapor concentrations, shown in
the previous two columns. The expected accuracy of L values is about one order of magnitude.

following section. The partial pressure of the osmotic agent increases with time, because it must
keep up with the increasing Laplace pressure inside
the shrinking bubble. During this stage, the squared
particle size decreases nearly linearly with time.
This stage contributes most to the overall bubble
lifetime.
(3) Eventually, the partial pressure of the osmotic agent
becomes high enough that it condenses into a liquid. The saturated vapor pressure of n-perfluorobutane at 37C equals 3.2 atm, and the condensation
of the bubble into liquid may occur anytime after
reaching this point (see the insert in Fig. 3A). After
condensation, the particle size and compressibility
exhibit a sharp drop and the particle becomes invisible to ultrasound. The nucleation of a new phase
may require supersaturation, and the exact moment
of the condensation is somewhat uncertain.
Figure 4 shows the effects of the interfacial tension and
the blood systemic pressure on the kinetics of the dissolution of an n-C4F10-filled bubble. The higher interfacial
tension and blood pressure, the shorter is the bubble
lifetime. Characteristically, these parameters mostly affect the ultimate swelling of the bubble at Stage 1, and
not the slope of the r2 vs. t lines. Figure 5 illustrates the
effect of the initial mole fraction of osmotic agent in the
bubble. The X F value also mostly affects the swelling of
the bubble and not the slope of the r2 vs. t lines: at high
values of X F , the bubble swells in air; at low values, it
shrinks.
Although the effect of condensation is not significant for C4F10, it becomes more important for longerchain perfluoroalkanes. Figure 6 shows that for
n-C6F14, the condensation may already occur at Stage

1, a few seconds after injection. This problem may be

circumvented, however, by the reduction of the surface tension from 70 to 20 mN/m, which makes the
n-C6F14 bubbles capable to withstand the Laplace
pressure (Fig. 7).
Figure 8 shows the predicted decay of the ultrasound signal due to the bubble dissolution, under the
assumption that the ultrasound intensity scales as the
sixth power of the radius. After the initial swelling
stage, the decay has an overall exponential form. At
the late stages, however, it is somewhat less steep than

Fig. 4. Effect of interfacial tension and effective excess pressure on dissolution of n-C4F10-filled bubble at r 0 5 2.5 mm and
X F 5 1. (A) s 5 70 mN/m; p# * 5 10 4 Pa; (B) s 5 70 mN/m;
p# * 5 4300 Pa; (C) s 5 70 mN/m; p# * 5 0; (D) s 5 20 mN/m;
p# * 5 10 4 Pa; (E) s 5 20 mN/m; p# * 5 4300 Pa; (F) s 5 20
mN/m; p# * 5 0.

744

Ultrasound in Medicine and Biology

Fig. 5. Effect of initial volume fraction of osmotic agent X F on

dissolution of n-C4F10-filled bubble at r 0 5 2.5 mm; s 5 70
mN/m; p# * 5 4300 Pa.

exponential. The duration of the signal increases in the

order: C2F6 , C4F10 , C5F12 , C6F14, with the
concomitant decrease in the Ostwald coefficient.
Very similar results were recently obtained by
Van Liew and Burkard (1995b). The authors reproduced Stages 1 and 2 of dissolution, but did not
consider the possibility that the osmotic agent may
condense.

Fig. 6. Dissolution of n-C6F14-filled bubble at X F 5 0.2, s 5

70 mN/m; p# * 5 4300 Pa; r 0 5 2.5 mm. The condensation into
liquid may occur already at Stage 1 of dissolution. For visual
purposes, the duration of Stage 1 is exaggerated.

Volume 24, Number 5, 1998

Fig. 7. Dissolution of n-C6F14-filled bubble at X F 5 0.2; s 5

20 mN/m; p# * 5 4300 Pa; r 0 5 2.5 mm. Reducing interfacial
tension shifts the condensation point to larger times.

ANALYTIC EXPRESSIONS FOR THE BUBBLE

LIFETIME
It has been shown that the second stage of the
bubble dissolution contributes most to the lifetime, if the
condition L F ,, L A is met and the effect of condensation
is negligible. One may analytically estimate the bubble
lifetime by considering this stage of dissolution. The key
idea is that the equilibration of the microbubble with air
is much faster than with osmotic agent, so the chemical
potential of air inside the bubble is almost equal to that

Fig. 8. Ultrasound scattering intensity, arbitrary units vs. time

(in s) for C2F6, n-C4F10, n-C5 F12 and n-C6F14-filled bubbles at
p# * 5 4300 Pa; r 0 5 2.5 mm; s 5 20 mN/m; X F 5 0.2. The
data are shown for the Rayleigh scattering limit, where the
scattering intensity is proportional to the sixth power of radius.

Microbubble ultrasound contrast A. KABALNOV et al.

in the lungs. Formally, this statement is equivalent to

setting the air mass flow, eqn (11) to zero:
3dLA
~ A 2 r 3! < 0
r2

(15)

Converting back to physical variables, these conditions

are equivalent to:
p A 5 p atm

(17)

and
p F 5 p blood 1

2s
.
r

(18)

This means that the partial pressure of air establishes at

nearly 1 atm to meet the condition of chemical equilibrium with air in the lungs, whereas the systemic blood
pressure and the Laplace pressure are counterbalanced
by the partial pressure of the osmotic agent alone. These
features are seen clearly in the computer simulations
(Figs. 3A and 3B). If one neglects the diffusion of osmotic agent out of the bubble in Stage 1 of dissolution,
one finds that at Stage 1 the bubble swells to the new
relative radius r , so that the following condition is met:
X F~1 1 n 1 m ! 5 nr 3 1 mr 2.

(19)

If the bubble initially contained a pure osmotic agent

(i.e., X F 5 1), it will swell in air; if X F , 1 it may swell
or shrink, in dependence of the n and m values (Fig. 4.)
The kinetics of dissolution at Stage 2 is determined
by eqns (10) and (16). After some algebra, one obtains
the following growth rate:

S D
dr
dt

21

1 2 r 1 3 jr 2
52
3L F 1 1 jr

(20)

where j 5 n/m. This equation can be integrated, with the

following expression for the bubble lifetime:

t lifetime 5

1 ln~1 1 jr ! r 3 r 2
2 1
.
3L F
j2
j
2

The logarithm can be expanded in series:

(21)

(22)

r 20
r 2
3D FL F

(23)

or, in physical units:

t lifetime <

which means that A 5 r . Substituting this into eqn (9),

one gets:
(16)

1 2
r
3L F

t lifetime <

F 5 mr 2 1 nr 3.

745

where r is determined by eqn (19). Note that eqn (19) is

cubic with respect to r and the explicit solution vs. r and
the subsequent substitution to eqns (21) or (23) is possible, but the final formula will be very bulky.
EFFECTS OF OXYGEN METABOLISM AND
SYSTEMIC BLOOD PRESSURE VARIATIONS
To date, air has been considered as a pseudocomponent and the fact that oxygen in the bubbles is consumed by living tissues was completely neglected. In a
more rigorous analysis, one must consider the mass
transfer of oxygen, nitrogen and carbon dioxide generated in the body separately. Indeed, only the tension of
nitrogen in the bloodstream is the same as in the atmosphere. Due to metabolism, the tension of oxygen in the
blood is lower than in atmosphere and varies between
about 40 and 100 mmHg among the different parts of the
bloodstream (Fig. 9A). Conversely, the tension of carbon
dioxide in the blood is about 40 mmHg and is larger than
in atmosphere. One may also expect some difference
between the partial pressure of water vapors in the bloodstream and in the atmosphere, e.g., due to the difference
in the body temperature and the atmospheric temperature. All these gases have high Ostwald coefficients and
equilibrate with the bubble within the first several seconds, so partial pressures of these gases in the bubble
must be essentially equal to their tensions in the blood.
The partial pressure of the osmotic agent must now
compensate not only for the systemic blood pressure and
Laplace pressure, but also for the excess partial pressures
of O2, CO2 and H2O in the atmosphere, in comparison
with the blood (Van Liew and Burkard 1995a):
pF 5

2s
1 p blood 1 Dp O2 1 Dp CO2 1 Dp H2O
r

(24)

where Dp O 2, Dp CO 2 and Dp H 2O are the excess partial

pressures of oxygen, carbon dioxide and water in atmosphere, in comparison with the blood.
Figure 9A shows that both the systemic blood pressure and the oxygen consumption are variable among the
different parts of the bloodstream. Thus, the relatively
high systemic pressure (120 80 mmHg) is applied to the
bubbles during a fraction of their overall circulation,

746

Ultrasound in Medicine and Biology

Volume 24, Number 5, 1998

1 Dp CO 2 1 Dp H 2O , shown in Fig. 9B. Although the

relative fluctuations in p* are large: p* ' 4300 6
10,000 Pa, the relative fluctuations in the partial pressure of the osmotic agent are less significant, because the
dominant contribution in p F is the Laplace pressure term
(Fig. 9B). This allows one to neglect the fluctuation of
the bubble size during dissolution and use an average
value of p*, which for humans is expected to be close to
p# * 5 4300 Pa (Fig. 9B). The whole analysis of the
previous section is therefore applicable, with the old n
value being replaced on n * 5 p# */p atm ' 0.043, the two
main equations becoming:
X F~1 1 n * 1 m ! 5 n * r 3 1 mr 2

(25)

and

t lifetime 5

1 ln~1 1 j * r !
r
3 r 2
2
1
2
3L F
~ j *!
j*
2

(26)

where j* 5 n*/m. Note that eqns (25) and (26) are

analytic and represent an improvement over Van Liew
and Burkard results, which are purely numerical.
CONCLUSIONS AND FINAL REMARKS

Fig. 9. (A) Bubble that contains the mixture of gases traverses

various parts of circulation. PA 5 pulmonary artery; PB 5
pulmonary vessel bed; PV 5 pulmonary vein; SA 5 systemic
artery; SB 5 systemic small vessel bed; SV 5 system vein.
Time spent in each part is arbitrary set at 6 s. Here p blood is the
systemic blood pressure. The Dp values are the excess partial
pressures in the atmosphere, in comparison with the tensions in
blood. (Reproduced, with some simplifications, from Van Liew
and Burkard (1995b). (B) Variations in effective pressure p*
and in partial pressure of osmotic agent in bubble p F as the
bubble passes through the bloodstream at r 0 5 2.5 mm and
s 5 70 mN/m.

when they pass through the systemic artery. Similarly,

the blood is substantially depleted in oxygen only when
passing through the systemic vein, pulmonary artery and
systemic small vessel bed. All these variations can be
summed into an effective pressure p* 5 p blood 1 Dp O 2

The problem of dissolution of a bubble in the bloodstream has been considered. The bubble is assumed to
contain a sparingly water-soluble osmotic agent and air.
The dissolution of the bubble can be separated into three
definite stages:
In Stage 1, the bubble quickly swells or shrinks to a
new radius. The swelling ratio depends on the surface
tension s, the systemic blood pressure, the level of
oxygen metabolism and the initial mole fraction of the
osmotic agent in the bubble X F , The larger X F , and the
smaller p# * and s, the bigger is the swelling ratio r . The
functional relationship between them is determined by
eqn (25).
In Stage 2, the osmotic agent slowly diffuses from
the bubble. At this stage, the partial pressure of the
osmotic agent counterbalances the Laplace pressure, systemic blood pressure, as well as the oxygen, carbon
dioxide and water vapor difference pressures. The partial
pressure of osmotic agent increases as the radius of the
bubble decreases, to keep up with the increasing Laplace
pressure.
In Stage 3, the partial pressure of osmotic agent
becomes so high that the osmotic agent condenses. The
particle size and compressibility of the bubble exhibit a
sharp drop, and the bubble becomes invisible to ultrasound. The particle size at which the condensation occurs

Microbubble ultrasound contrast A. KABALNOV et al.

can be markedly decreased by diminishing the surface

tension.
If the condensation effect is negligible, the time of
complete dissolution can be evaluated with eqn (26). The
squared radius is predicted to nearly linearly decrease
with time, with the slope proportional to the Ostwald
coefficient and diffusivity of the osmotic agent. The
lifetime of the bubble is, however, not proportional to the
squared initial radius, because the initial swelling ratio
depends on r 0 .
In order to prolong the lifetime of d 5 5-mm
bubbles in the bloodstream from a fraction of a second to
minutes, an osmotic agent must combine a low Ostwald
coefficient value of ; , 1024 and a high saturated vapor
pressure at body temperature of . ; 0.3 atm 5 3 3 104
Pa. These requirements make the scope of possible candidates rather narrow. No polar gases (such as ammonia,
carbon dioxide or other water soluble gases), or inert
gases, such as Ne or Ar, satisfy these conditions (Wilhelm et al. 1977). Even hydrocarbons are unable to meet
these criteria.
APPENDIX
Selection of parameters
Particle size. All calculations were made for the
initial radius r 0 5 2.5 mm, which is in the middle of the
acceptable range for the ultrasound contrast media. The
system was assumed to be monodisperse.
Surface tension. Surface tension s at the bubblewater interface was assumed to fall within the range
70 20 mN/m. The first value corresponds to the surface
tension at the air-water interface at 37C and represents
the absolute upper limit in this experimental set-up.
Because of the plasma protein adsorption, the tension of
a naked bubble, not stabilized by surfactants, is probably close to 50 mN/m (Van Liew and Burkard 1995b).
If microbubbles are covered by surfactant layer, the
tension can be decreased to, at most, ; 20 mN/m for
hydrocarbon surfactants, and ; 16 mN/m for fluorocarbon surfactants (Laughlin 1994). Further decrease is
possible only due to non-equilibrium adsorption effects
(Pison et al. 1996), which are beyond the scope of this
study. Accordingly, the dimensionless Laplace pressure
parameter m varied between 0.56 and 0.16.
Effective excess pressure. This value is determined
as the sum of the excess systemic blood pressure p blood ,
and the excess oxygen Dp O 2, carbon dioxide Dp CO 2 and
water Dp H 2O vapor partial pressures in the atmosphere
vs. the bloodstream:
p* 5 p blood 1 Dp O2 1 Dp CO2 1 Dp H2O.

(27)

747

The sign of the excess values is determined in such a way

that lower tension in the blood vs. the atmospheric partial
pressure is counted as a positive excess; accordingly,
Dp O 2 . 0 and Dp CO 2 , 0. In the analysis of the
variations in p blood , Dp O 2 and Dp CO 2 during the bubble
circulation, we rely on the article by Van Liew and
Burkard (1995b). The Dp O 2 and Dp CO 2 values vary as
the blood passes through the systemic vein, pulmonary
artery, pulmonary vessel bed, pulmonary vein, systemic
artery and systemic small vessel bed, as shown in Fig. 9A.
Because the body temperature is slightly higher than the
temperature of the surrounding medium, the excess partial water pressure is expected to be negative: Dp H 2O ,
0. The vapor pressure inside the bubble is saturated at
37C (6.28 kPa), while the partial water pressure in the
atmosphere can vary from 0 (zero humidity) to about
2.34 kPa (saturated vapor pressure at 20C). Accordingly, Dp H 2O can vary from 26.28 to 23.94 kPa; the
value of 4 kPa was assumed in the calculation. Time
averaging of p* yields the value of p# * 5 4.3 kPa.
Accordingly, in most simulations, the dimensionless effective pressure parameter n* was assumed to be equal to
n* 5 0.043. The expected scatter for this value is quite
significant: n* 5 0.043 6 0.1.
From the previous paragraph, it is evident that the
estimate for p# * relies on many assumptions. However, it
provides only a small correction to the dissolution time,
because the leading term is still the Laplace pressure
term, cf. Fig. 4.
Saturated vapor pressures. Saturated vapor pressure of the perfluoroalkanes studied were tabulated by
Reed (1964). The values are shown in Table 1.
Diffusion coefficients. Diffusion coefficients of fluorocarbons in water at 25C were evaluated with the
Hayduk-Laudie equation from the molecular volumes of
fluorocarbons (Hayduk and Laudie 1974):
D 5 13.26 10 25 3 V 20.589
.
m

(28)

For diffusion coefficients of nitrogen and oxygen in

water, the literature data are available (Lide 1991). The
diffusion coefficients in the blood are expected to be of
the same order of magnitude as those in water. The
values refer to 20 25C and no attempts to make a
correction for 37C is made.
Water solubilities and Ostwald coefficients. The
data are shown in Table 1. Aqueous solubilities of some
liquid perfluoroalkanes at 25C, including n-perfluoropentane and n-perfluorohexane c water were estimated by
Kabalnov et al. (1990). On the basis of homologous
series studies, it was found that the solubility decreases
by a factor of 8 per each CF2-group added. This allows

748

Ultrasound in Medicine and Biology

one to evaluate the solubility in water of (metastable)

liquid perfluoroethane and perfluorobutane at 25C by
approximating the homologous series data. The concentration of the equilibrium saturated vapor at 25C was
estimated with the ideal gas equation:
C gas 5

p sat
.
RT

(29)

Ostwald coefficients were then evaluated with eqn (3).

The temperature dependence of the Ostwald coefficients
(25C vs. 37C) is neglected. The expected accuracy is
about one order of magnitude.
Notation. Throughout the article, the SI system is
used, except for the pressure units, which are shown in
Pa, and, occasionally, in mmHg, and atmospheres. The
relationship between these units is: 1 3 105 Pa 5 1
atm 5 760 mmHg.
r Particle radius
t Time
r 0 Initial radius (at t 5 0)
s Surface tension
R Gas constant
T Absolute temperature
c water Solubility of the gas in
water (mole/m3)
C gas Saturation concentration in
the vapor phase, (mole/m3)
L 5 c water /C gas Ostwald coefficient
L F , L A Ostwald coefficient of the
osmotic agent and air,
respectively
D F , D A Diffusion coefficient of the
osmotic agent and the air in
blood, respectively
V m Molar volume of the liquid
osmotic agent
p sat Saturated vapor pressure of
the osmotic agent at the
given temperature
p atm Atmospheric pressure
(51 3 105 Pa)
p blood Systemic blood pressure
C F , C A Concentrations of the
osmotic agent and the air in
the bubble (mole/m3),
respectively
c F , c A Concentrations of the
osmotic agent and the air in
the blood (mole/m3),
respectively

Volume 24, Number 5, 1998

J F , J A Mass fluxes of the osmotic

agent and the air from the
bubble (mole/s),
respectively
p F , p A Partial pressures of the
osmotic agent and the air in
the bubble, respectively
2s
m 5
patm r Dimensionless Laplace
pressure
pblood
n 5
patm Dimensionless systemic
pressure
r
r 5
r0 Dimensionless radius
DA
d 5
DF Ratio of diffusion
coefficients
CA RT
xA 5
,
patm
CF RT Dimensionless
xF 5
patm concentration of the air and
osmotic agent in the bubble
DF
t 5 2 t
Dimensionless time
r0
x Fr 3 5 F
x Ar 3 5 A
j 5 n/m
t lifetime Time of complete
dissolution
X F Mole fraction of the
osmotic agent in the
mixture with air inside the
bubble at t 5 0
r Swelling ratio, which is the
ratio of the particle radius
after the first, quick stage
of dissolution to the initial
size r 0
Dp O 2 Difference between the
partial pressure of oxygen
in the atmosphere and the
oxygen tension in the blood
Dp CO 2 Difference between the
partial pressure of CO2 in
the atmosphere and the
CO2 tension in the blood
Dp H 2O Difference between the
partial pressure of water in
the atmosphere and in the
bubble

Microbubble ultrasound contrast A. KABALNOV et al.

p* 5 p blood 1 Dp O 2 1
Dp CO 2 1 Dp H 2O Effective excess pressure
p# * Time-average effective
excess pressure
n * 5 p# */p atm
j * 5 n */ m
I Scattered intensity
I 0 Incident intensity
n Number density of
scattering particles
V Scattering volume
k Wave number
u Scattering angle
d Distance from scatterers
g c Compressibility term
g d Density term
k s , k m Compressibilities of the
scatterer and the medium,
respectively
r s , r m Densities of the scatterer
and the medium,
respectively
REFERENCES
Balen FG, Allen CM, Lees WR. Ultrasound contrast agents. Clin
Radiol 1994;49:77 82.
Bernstein H, Straub JA, Brush HT, Wing Richard E. Microencapsulated fluorinated gases for use as imaging agents. US Patent No.
5611344, 1997.
Burkard ME, Van Liew HD. Simulation of exchanges of multiple gases
in bubbles in the body. Resp Physiol 1994;95:131145.
Correas JM, Quay SD. EchoGen emulsion: A new ultrasound contrast
agent based on phase shift colloids. Clin Radiol 1996;51(Suppl.
1):1114.
Crank J. The Mathematics of Diffusion. Oxford; Clarendon Press,
1975.
de Jong N. Acoustic Properties of Ultrasound Contrast Agents, Ph.D.
thesis. Rotterdam, The Netherlands; Erasmus University, 1993.
Epstein PS., Plesset, MS. On the stability of gas bubbles in liquid-gas
solutions. J Chem Phys 1950;18:15051509.
Evans DF, Wennerstrom H. Colloidal domain. New York: VCH Publishers, 1994.
Fritzsch T, Schlief R. Future prospects for echo-enhancing agents. Clin
Radiol 1996;51(Suppl. 1):56 58.
Goldberg BB, Liu J-B, Forsberg F. Ultrasound contrast agents: A
review. Ultrasound Med Biol 1994;20:319.
Gramiak R, Shah PM. Echocardiography of the aortic root. Invest
Radiol 1968;3:356 366.
Hayduk W, Laudie H. Prediction of diffusion coefficients for nonelectrolytes in dilute aqueous solutions. AIChE J 1974;20:611 615.
Kabalnov AS, Pertsov AV, Shchukin ED. Ostwald ripening in twocomponent disperse phase systems: application to emulsion stability. Coll Surf 1987;24:19 32.

749

Kabalnov AS, Makarov KN, Shcherbakova OV. Solubility of fluorocarbon in water as a key parameter determining fluorocarbon emulsion stability. J Fluor Chem 1990;50:271284.
Kabalnov A, Weers JG, Schutt EA. Gas emulsions stabilized with
fluorinated ethers having low ostwald coefficients. US Patent Application No. 08/479,621, 1995.
Laughlin RG. The Aqueous Phase Behavior of Surfactants. San Diego:
Academic Press, 1996.
Lawson DD, Moacanin J, Scherer KV Jr, Terranova TF, Ingham JD.
Methods for the estimation of vapor properties and oxygen solubilities of fluorochemicals for possible application in artificial blood
formulations. J Fluor Chem 1978;12:221236.
Lide DR. CRC Handbook of Chemistry and Physics. Boca Raton: CRC
Press, 1991.
Lohrmann R. Gas-filled microspheres with fluorine-containing shells.
US Patent No 5562893, 1996.
Mattrey RF, Wrigley R, Steinbach GC, Schutt EG, Evitts DP. Gas
emulsions as ultrasound contrast agents: Preliminary results in
rabbits and dogs. Invest Radiol 1994;29;S139 S141.
Meltzer RS, Tickner EG, Popp RL. Why do the lungs clear ultrasonic
contrast? Ultrasound Med Biol 1980;6:263269.
Morse PM, Ingard, KU. Theoretical Acoustics. New York: McGraw
Hill, 1968.
PCR Research Chemicals Catalog. Gainesville, FL: PCR Inc., 1992.
Pison U, Herold R, Schurch S. The pulmonary surfactant system:
Biological functions, components, physico-chemical properties and
alterations during lung disease. Coll Surf A 1996;114:165184.
Porter TR, Xie F, Kricsfeld D, Armbruster RW. Improved myocardial
contrast with second harmonic transient ultrasound response imaging in humans using intravenous perfluorocarbon-exposed sonicated dextrose albumin. J Am Coll Clin Cardiol 1996;27:1497
1501.
Quay SC. Methods for selecting and using gases as ultrasound contrast
media. US Patent No. 5,393,524, 1995.
Reed TM. Physical chemistry of fluorocarbons. In: Simons JH, ed.
Fluorine chemistry, vol. 5. New York: Academic Press, 1964:
.
Schlief R. Developments in echo-enhancing agents. Clin Radiol 1996;
51(Suppl. 1):57.
Schneider M, Yan F, Grenier P, Puginier J, Barrau, M-B. Long-lasting
aqueous dispersions or suspensions of pressure-resistant gas-filled
microvesicles and methods for preparation thereof. US Patent No.
5,413,774, 1995.
Schutt EA, Evitts DP, Kinner RA, Weers JG, Anderson CD. Stabilized
microbubble compositions for ultrasound. US Patent No. 5639443,
1997.
Unger EC, Matsunaga T, Ramaswami V, Yellowhair D, Wu G. Methods of preparing gas and gaseous precursor-filled microspheres.
International Patent Application No. PCT/US94/05792, 1994.
Van Liew HD, Burkard ME. Bubbles in circulating blood: Stabilization
and simulations of cyclic changes of size and content. J Appl
Physiol 1995a;79:1379 1385.
Van Liew HD, Burkard ME. Behavior of bubbles of slowly permeating
gas used for ultrasonic imaging contrast. Invest Radiol 1995b;30:
315321.
Wilhelm E, Battino R, Wilcock RJ. Low-pressure solubility of gases in
liquid water. Chem Rev 1977;77:219 262.
Yang W-J. Dynamics of gas bubbles in whole blood and plasma.
J Biomech 1971;4:119 125.